摘要
针对基于成像式光电容积描记术(IPPG)检测人体生命体征会产生运动伪影的问题,提出一种两阶段运动伪影消除的心率检测算法。所提算法使用基于特征点的人脸跟踪算法S-KLT(Scale invariant feature transformKanade Lucas Tomasi)稳定感兴趣区域(ROI),采用基于自相关的自适应滤波器抑制残余运动伪影。首先,采集视频,使用S-KLT追踪人脸ROI内的特征点,提取ROI内原始的IPPG脉搏波信号;其次,利用所提基于自相关的自适应信号滤波器对原始的IPPG信号进行噪声抑制,根据峰值检测算法计算心率(HR)值。实验结果表明,与腕式电子血压计实际测得的HR值相比,所提算法的均方根误差(RMSE)、平均错误率(EER)分别为2.98%、2.90%,符合美国医疗器械促进协会(AAMI)的标准,可以有效抑制运动伪影对HR值准确度的影响,为远程医疗和日常的健康检测提供了新的解决方法。
Aiming at the problem of motion artifact based on Imaging PhotoPlethysmoGraphy(IPPG)in the detection of human vital signs,a two-stage heart rate detection algorithm with motion artifact elimination was proposed.A feature-point based Scale invariant feature transform-Kanade Lucas Tomasi(S-KLT)was used to stabilize the Region Of Interest(ROI),and an adaptive filter based on autocorrelation was used to suppress residual motion artifacts.Firstly,the video was collected,and the S-KLT was used to track the feature points in the ROI to extract the original IPPG pulse wave signal in ROI.Secondly,the proposed adaptive signal filter based on autocorrelation was used to suppress the noise of the original IPPG signal,and the Heart Rate(HR)value was calculated according to the peak detection algorithm.Experimental results show that compared with the actual HR value measured by wrist electronic sphygmomanometer,the Root Mean Square Error(RMSE)and Equal Error Rate(EER)of the proposed algorithm are 2.98%and 2.90%,respectively,which meet the standards of the Association for Advancement of Medical Instrumentation(AAMI),and effectively suppress the influence of motion artifacts on the accuracy of heart rate.It provides a new solution for telemedicine and daily health examination.
作者
李牧
李倩
柯熙政
陶启婷
LI Mu;LI Qian;KE Xizheng;TAO Qiting(School of Automation and Information Engineering,Xi’an University of Technology,Xi’an Shaanxi 710048,China)
出处
《计算机应用》
CSCD
北大核心
2023年第S01期333-339,共7页
journal of Computer Applications
关键词
成像式光电容积描记术
心率检测
运动伪影
人脸追踪算法
自适应信号降噪
Imaging PhotoPlethysmoGraphy(IPPG)
heart rate detection
motion artifact
face tracking algorithm
adaptive signal noise reduction
